Short-time pasteurization of milk - ACS Publications

Short-time pasteurization of milk. C. OLIN BALL, Owens-Illinois Can Company, Toledo, Ohio. Certain problems connected with short-time pasteurization o...
0 downloads 0 Views 3MB Size
Short-time pasteurization of milk C. OLIN BALL, Owens-Illinois Can Company, Toledo, Ohio

Certain problems connected with short-time pasteurization of milk have not been brought to a solution. The principles employed successfully in the scientific advancement of heat sterilization of canned foods are interpreted here so as to facilitate the application to milk pasteurization of a treatment analogous to that applied to canned foods. This interpretation is made in terms familiar to those acquainted with process calculation methods for canned foods. A manner of using this method to solve the problems associated with milk pasteurization is described, with emphasis on the explanation

of the high thermoduric bacteria counts in milk which has been adequately pasteurized by short-timemethods. The use of the process evaluation method in the scientific solution of pasteurization problems must be inaugurated by the accumulation of data that can be used in determining significant values in thermal death time relations for the pathogenic and thermoduric microorganisms associated with milk pasteurization. The manner of correlating this information with time-temperature curves for milk in the pasteurization process to evaluate the lethality of the process is illustrated by examples.

F

OR years after milk pasteurization became a commerical process, the idea was prevalent that pathogens in milk could not be destroyed a t a temperature below 60' C. (140' F.); in other words, 60' C. was considereda critical temperature that had to be attained if the dangerous microorganisms in milk were to be destroyed. Many believed that pasteurization could be satisfactorily accomplished only between 60' and 62.8' C. (145" F.). Later the idea gained acceptance that there is no critical lethal temperature, that any temperature high enough to have an unfavorable effect upon the growth and stamina of the bacteria is lethal, and that the bacteria will be destroyed if they are subjected to that temperature long enough. A corollary is the idea that, because less time is required to destroy bacteria a t high than a t low temperatures, advantages might be gained by pasteurizing milk above 62.8" C . . With the growth of this thought the principle of high-temperature short-time pasteurization began to gain support. Because of certain advantages of high-short pasteurization of milk, the long-hold form of pasteurization, exemplified by the customary exposure of market milk for 30 minutes at 61.1' to 62.8' C. (142' to 145" F.), has been yielding ground slowly .but steadily to short-time methods in recent years. For market milk the minimum exposure accepted in highshort pasteurization is 71.1' C. (160' F.) for 15 seconds. An exposure commonly used is 71.7' C. (161' F.) for 16 seconds. The growing use of the short-time form of pasteurization brings the necessity of increased stress on accurate time and temperature control, because a slight error in either time or temperature may have a much more serious effect in a shorttime process than in a long-hold process. Difficulty in establishing dependable temperature and time control has been one of the primary retarding factors to increased use of shorttime milk pasteurization. Even with perfect manipulation of the process, one disturbing condition has occasioned much discussion. The counts of thermoduric bacteria are relatively high in milk which, according to standard criteria, has been adequately pasteurized by short-hold methods. A purpose of this paper is to

supply a scientific explanation of this phenomenon and a t the same time to present an approach to a Scientific attack upon the general question of lethal value of milk pasteurizing processes. To accomplish this purpose we turn to the principles of thermal death time that are applied in mathematical methods of evaluating canned foods processes. An interpretation of these principles is necessary before a new application can be made. This interpretation can be presented most satisfactorily with the use of terms that are commonly employed in process calculation methods for canned foods; and although this material may appear to be foreign to the subject, it is an essential part of this presentation.

Evaluation of canned foods processes

FUNDAMENTAL PRINCIPLE. There are products through which heat penetrates so slowly that, in containers of large size, the center does not reach the temperature of the retort during a process of practical length. After heat penetration tests showed this to be true, i t became clear that canned food processes could not be standardized from the standpoint of sterilizing value solely on the basis of the length of time the center of the can is held a t retort temperature. The entire period during which the center of the can is a t temperatures which are lethal to bacteria would have to be considered. BASISOF PROCEDURE. The key to evaluating a process for sterilizing value is found in the heat penetration curve and the thermal death time curve (d,b, 4,6-9,13). Such a curve shows that the spores of a microorganism are destroyed in 10 minutes a t 115.5' C. (240' F.) or in 360 minutes a t 100' C. (212' F.). In terms of lethal effect, it is thirty-six times more intense a t 115.5" than a t 100' C. I n other words, the intensity of the lethal effect is inversely proportional to the length of time required to destroy the organism. For a can of corn a t 121' C. (250' F.), for instance, the effective energy which is in the act of destroying this spoilage organism a t the point of maximum lag (point of slowest heating) in the can is thirty-six times as great when the maximum lag point is a t 115.5"as when it is a t 100"C. 71

sa

Vol. 35, No. I

To evaluate the process, we must apply a sterilizing, or lethal, value to each temperature through which the maxinium lag point of the can passes and add these together in such a way as to give full value to the length of time the ~11ax1mum lag point is held a t each temperature. During the time the temperature a t the maximum lag point of the can is rising and fdling, the point IS a t a given temperature only an infinitesimal length of time. Thus, to add the lethal effects for a11 temperatures, calculus must be employed.

(,XI’=

il

CIL

-$=

1

R E C I P R O C A L OF T I M E IN MINUTES

The object in prOCeRLCUL4TION OF STERIEIIIIYG \‘ALEE. essing canned foods is the attainment of sterility with respect to the most resistant microorganism present which mould bring about spoilage. Therefore the problem of determining %hetinie necessary to process a canned food consists of determining the time necessary to produce this sterility within the cans, A method of eomput,ing this time is described by Bigelow et al. ( 6 ) . This is the foundation of the calculation method described by Ball ( 2 , 3, 4) and further developed by Olson et aE. (If, 16, 22, IS), Each minute section of the heating and cooling curves for the maximum lag point of a can, corresponding t o a temperature that has lethal effect on spoilage organisms, is said to have a lethal rate value. This value, for any section, i s the reciprocal of the number of minutes r e q u i d to destroy ala spores of the organism a t the temperature represented by the midpoint of the section, under the condition obtaining within the food, times the length of the time period represented by the section 0.05 of t h e c u r ~ 7 c . 0.04 These values c g 0.03 are plotted against time, 0.02 talien from the 0.01 abscissas of the 0 r e s p e c t i v e sec0 10 20 30 40 Lion mid-points TIME IN MINUTES of t h e t i m e Figure 2. Lethalf& Curves temperature heating and eooling curves. GRAPHICAL TRIAL-AND-ERRQE METHOD. The method described by Bigelow is primarily graphical; the mechanics are illustrated in Figures I and 2. Figure B shows the timetemperature curves of the center of a can during heating and cooling, and the thermal death time curve of an organism. The curves are plotted side by side so that the same scale of degreea applies to all. The abscissas represent time; z’ is time as determined in a heat penetration test of the can of food, 5’’ is time as determined in a laboratory test of the re5istance of the microorganisms to heat. For the thermal 9

: 3

death time curve, 2’’ i s a scale of the reciprocals of the normal abscissa scale, z’, The broken lines joig., values on the b~ scale of processing time to values on the z” scale of reciprocals of heat resistance t’ime, which are plotted against each other in Figure 2 , i[n this new curve, valuea from the .?; scale of Figure 1 are represented as abscis,sas, and ualues from the 5’‘ scale as ordinates. The &reabeneath the lethality curve (Figure 2 ) is equal to unity if the process represented by the heating and cooling curves of Figure I is just suficient to destroy ail spores of the organism; and the time, s (Figure I), is then the length of process in minutes. This is strictly a trial-and-error met,hod of determining the length of a process necessary t o sterilize a can of food. If the area beneath the lethality curve is either greater or less than unity, the process of solution must be repeated in order to determine the required length of process. ~ A ~ Msz.r~ous---In ~ ~ morea recently ~ developed ~ ~ methods (2, 3, dI7 the trial-and-error feature of the determination of sterilizing value is eliminated, A determination is nzade by direct mathenaatical calculations based on estahlished formulas. The formulas are developed through the integration of lethal rate values assigned to the successive points of the heating and cooling curves of a can. To show g r ~ p h ~ ~ ahow l l y this development is accomplished, B strip i s taken along the heating and cooling curves which has B constant width of unity, measured in the direction of ordinates.

This strip is divkled into elements having infinitesimal horizontal width, ds. The elements extend across the strip in the direction of ordinates; therefore each element is of unit length ”2) - (y’ dz, which eqglals dz and i t s area is (y (Figure 3). Even though these elements have the same size, they do not all have the same value. Like gaseous molecules, each may be said to have its own weight. Just as the chemical atomic might of an element is indicatcd by the position of the element in the periodic table, the atomic weight (lethal rate value) of each of these elements is indicated by the position of the element with Tespeet to the temperature scale. Figure 3 shows a heating and a cooling curve, symbolically enclosed within a strip, the vertical distance across tlie strip is 4 a t all points. An element of the strip having an infinttesimal area, dz, is also indicated, Either the heating or the cooling curve passes through the exact center of every element. Each element may be said to have a temperature value, taken from the temperature scale along the left-hand border of Figure 3 a t the elevation of the center point of the element.

+-

~

c

lanuary, 1943

INDUSTRIAL AND ENGINEERING CHEMISTRY

73

ture portions of the curve. The complete summation can be accomplished only by calculus. Calculus requires the use of mathematical equations for those curves. It was found that satisfactory equations could be developed for the heating, cooling, and thermal death time curves, on the basis that the thermal death time curve, the portion of the heating curve in which we are interested, and all of the cooling curve except the first part can be represented as straight lines on semilogarithmic coordinate paper, and that the first part of the cooling curve can be represented as a hyperbola. A heating curve for a can and a thermal death time curve plotted on semilog paper are shown on Figures 4 and 5, respectively. Symbols representing some of the numerical properties, which are used in the calculation of processes, are shown. The standard symbol for designating the slope value of a heating curve isfh, expressed as minutes, and of a thermal death time curve, z, expressed as O F. The thermal death time curve on Figure 5 is considered the ideal destruction curve for Cl. botulinum. After the equations have been derived, Courtesy, York Ice Machinery Corporation the area of an elementary portion of the Milk Pasteurizer of Short-Hold Type, Plate Construction, with a Caoacitv . * strip enclosing each curve is expressed of 8000 Pounds of Milk per Hour in terms of the equation of that curve. Each of these expressions (one for the heating curve, one for the hyperbolic part of the cooling curve, and one for the semilogarithmic The temperature value of the element is significant in that part of the cooling curve) is multiplied algebraically by an it enables one to determine the "weight" of the element, which is dependent not only upon the area, dx, but also upon expression representing lethal rate value derived from the equation of the thermal death time curve. The resultant the lethal rate value corresponding to the temperature of the element. As explained previously, the lethal rate value is the expressions represent the weights of the area elements in the reciprocal of the destruction time of a microorganism at a strips enclosing the respective curves. given temperature. This value is readily obtained with the assistance of the thermal death time curve for the microorgan240 ism, which shows the destruction time for each temperature. If, to each element is assigned a weight, dependent upon the lethal rate value corresponding to the temperature value of 2 5 the element, it is obvious that a summation of the weights of z all elements on the heating and cooling curves of a can will + give the sterilizing value of the process represented by those 245 curves. The weight of any portion of the strip which encloses i E the heating curve is expressed as the product of the lethal 8 240 rate value and the area of the portion of the strip being considered-for example, 0.1 X dz, for an element having an area dx when the lethal rate value is 0.1. To assist in developing a concept of this summation operation, we may assume that the heating curve in Figure 3 is 200 extended to reach retort temperature, 240" F., and held a t that temperature for a period of time. Along the horizontal portion of the curve (at 240") all elements of equal size have I50 equal weight. If we base a process upon a thermal death time curve of a microorganism that is destroyed in 10 minutes a t I I I I I I I 1 0 20 40 SO 80 100 120 t40 240" F., the weight of each of these elements is 0.1 dx. The T'IME IN MINUTES area of all the elements on a portion of the heating curve Figure 4. Heat Penetration Curve (Retort extending over a distance represented by one minute of time Temperafure Equals 250' F.) is unity, or 1. The weight, therefore, of such a section of the strip is 0.1 X 1 = 0.1. It is clear that the weight, or sterilizing value, of the portion of the process during which the maxiThese expressions are integrated between the point a t mum lag point of the can is a t 240" F. is 0.1 times the length which the heating curve reaches a point 80" F. below retort of that portion of the process expressed in minutes. temperature and the point a t which the cooling curve reaches The summation of "weights" of area elements along porthe same temperature. The integration leads to an equation tions of the curves which represent changing temperature is in which the total "weight", A , of the strip enclosing the more complicated than the summation for constant-tempera-

~ e ~ and t cooling ~ ~ gcurves is represented by the algebraic express~on,f,&/t, where j~ ~ e ~ r e s the e n s~ l ~ of p the heating and cooling curves, t is the number of minutm required $0 ~ ~ s tthe r oorganism ~ at the hghcst teesnper~tureattairbed by the rnaxirnum lag point of the can in the proces~,and is aka ~ ~ ~ constant. ~ ~ ~When 5 Ar =y I, the equdion represents the Cipndition of sterihty. The $OlUtiQIlsOf prOW9sing probHems are based upon this equation, which is solved with the use of B table giving valrres of C corresponding to different d u e s St" Various factors.

e

ment of processes because it must be assumed &laatGE. botuld-, w ~ r aof maxirnnm heat resistance, as found by Esty an might be present in food that i s canned, &CfhPIBB of these diEoulties, atteinpts have been made to hardeSs Q r ~ a k a G~f sCO%a8taat ~~ ~ ~ ~ ~ r a ~ Which t ~ r i s ~ ~ c $ might give results in laborstory tests simiIar to those that Would be Obf23&n@d W i t h @I+ ~~~~~~~~,~~~ $ % a d dthe h 3 k r have constant heat resistance. No organism entirely satisfactory for this purpose has been found, The one most ~ r ~ ~ ~ ~hspi g ~ c a ~ used, which, on the basis of general ~ ~ ~ r p p h traits, been thought to be more like C1. botulinum than any other nontoxic organism, is a putrefrac.i;ive&yIBePobe, designated by National Canners Associatiou as No. 3099 (American Type Culture Collection : ~ Z o s % r ~ ~ ~ u ~No. 7955) This orgaiiism, however, is now declared to be fundamentally different from Cl. ~ o ~ (13, ~ g7). l ~ ~ ~ m But why, after heat resistance of has been determined, is it necessary to ti^^^ ~ ~ ~ ~ rteats a t with o ~ y the organism? The re,asoxa ia that the organism varies not only between stmi given strain, according to the type of ganlisrn i s heated as well as to the type in which i t is cultured. The organisms must be studied in every differpent food in all of its variations, and the resistance not only at one ternpereture but at four or Eve different t e ~ ~ ~ e r a t umust r e s be determined so that the thermal death time curve applying to each given set of condit.ions will be est^^^^^^^^^^ A single culture of an organism may have many thermal death time curves which difler from one another not sady im absolute v a h e for w, given t ~ ~bad ahan ~ in slope: ~ r ~ (COmInOnly designated hg. #)? l4JXGYdhg &cS tha? food h?, 'Which the organism is steadied, The probleans that must be solved in order to establish thermal death time curves for C%.Botu& R U are ~ set forth c l e d y by TO'AIBS~PB~, Esty, and Bmel (br), who show why a complete thermal death time GUPV must be built up for each individual food, from ~ a ~ ~ r experimeiits with the C1. botulinum organism itself. ~~~~~~~~~~

~

a"

~~~~~~~~~

'khe development O f specific solution met~hsdsfrom this foB.pnsUh WBB EbCCOlnplished through a h l g ZiJK! inVOlVed pB"0Cem. Notable impr~ve~nents in the use of the forxada, resulting in simplification of calcdations, ~ R Wbeen made by Olson a& ai. (11, 18, 23). eat resistance ~

~ v ~ ~ BOT'CJLINUM. O 'h3rEletl~ death time,T01'heab; ke- ~ sistance, tests are made on cultures of pure strains of microort iEmpOrtaBlt in the problem Ob stO1'ihing ~ ~because this ~ oyganism ~ secretes ~ an extremely potent toxin. Thus, the first essential of every sterilizing process is that it be adequate to destroy the spores in this organism in the food. Although the spores are se%dom present in food prepared for canning, it must be assumed, because there is no quick way of detecting them, that the spores are always present. AlB known difficulties of heat resistance tests are encountered in working with the botulinus problem, 17-hereas not all. of the difficulties are met in dealing with certain o t h r species of bacteria. Therefore, we shall discuss heat rcsistaaco with particular reference t o protection of caniied foodr. against C1, botulinum, %"OS t ~ reasons o laboratories have avoided working with c8. ba%uhuma5 much as possible: There is a health hazard, and Ci. botuhurn has more than the usua,l degree of inconstancy in hea,tresistance of spores from d.if€erent cultures. No method is known of predicting heat rcsistance of spores of a given culture of this prganism. Therefore, if a culture is produced with the intention of using it in processing studies, the chance is great that the organism will have low resistance to heat, and that results will be misleading in the establish-

~

~ ~

~

~

~

n

i

~

~

~

~

,

~ ~

January, 1943

I N D U S T R I A L A N D E N G I N E E R I N G CHEMISTRY

Apparatus for Determination of Spore Destruction Rates (30)

FOOD-PHOSPHATE FACTOR. A practice followed to express the differences in resistance of spores in different media is to find the resistance of a particular strain of the organism in the food product in question and in buffered phosphate medium at only one temper9ture. The ratio, of the two numerical values in minutes has been known as the food-phosphate factor. On the assumption that this ratio is constant for all temperatures and thus that all curves for Cl. botulinum have the same slope, the food-phosphate factor has been used to establish a tentative thermal death time curve for the organism in the particular food product being studied. Discovery of the invalidity of this assumption revealed the weakness of the food-phosphate factor method of procedure in establishing processes for canned foods. The reason that organism 3679 is not a true substitute for C1. botulinum in heat resistance tests is based on a related principle-namely, that with a given food, the two organisms give thermal death time curves of different slopes (27). Thermal death time data that are truly applicable to Cl. botulinum, therefore, must be obtained in tests made with the C1. botulinum organism itself. EVOLUTION OF HEAT RESISTANCE METHODS.I n twenty years there has been an interesting evolution in methods of determining heat resistance of bacteria; and as this evolution has progressed, uncertainty has grown as to just what thermal death time is. The uncertainty is directly traceable to the complexity of those comparatively simple living organisms, the bacteria. The question we must answer is, “When are bacteria destroyed?” Present practice in determining heat resistance as applied to the evaluation of canned food processes includes methods which may be put into five general classes. Five different types of apparatus are employed variously in carrying out these methods. Two of these types of apparatus are used with all classes of methods; the other three have selective applications to the methods. Some of these methods and techniques are described in the literature ( 7 , 9 ,SO). There has been an evolution in thermal death time test

75

procedure, advanced by a growing realization of the extreme complexity of this problem. The trend appears to be away from the procedures that show absolute destruction points and toward those that reveal rates of destruction of bacteria in terms of numbers destroyed in set periods a t constant temperature, without proceeding to absolute destruction. This trend is based upon the principle that large numbers of units are capable of providing more consistent data than small numbers of units. In most rate-of-destruction tests bacteria are counted, whereas in tests designed to give absolute destruction points, containers, as a rule, are the units counted. Investigators feel that the most logical method for determining the destruction point of bacteria lies in determining rates of destruction by the plating method of testing. The results of tests to determine rate of destruction of spores are conveniently recorded by rate-of-destruction curves, which show a relation between time and numerical strength. The time is that of heating the bacteria a t a given temperature, and the numerical strength is the number of viable organisms per unit quantity of material. The normal rate of destruction of bacteria in a pure suspension a t a constant temperature is said to be that of a chemical reaction. Although the rate usually increases more than twofold (as does a monomolecular chemical reaction) for each 10’ C. rise of temperature, it seems safe to consider the normal rate-of-destruction curve to be a straight line on semilog paper (28). For the present, therefore, we shall disregard the possibility of having “sagging” or “bulging” curves, for which Rahn (20, 21) assumed various explanations. By definition, a reaction that proceeds a t a logarithmic rate never reaches a n end point; this means here that, if we heat a n infinite quantity of material containing an infinite number of bacteria, we never destroy every bacterium, or if we heat a finite quantity of material containing a finite number of bacteria, we never completely destroy the last bacterium. I n the broadest sense, therefore, theoretically there is no thermal death timethat is, a heating period a t the end of which all bacteria are dead.

(Above) Thermal Death Time Cans (2*/2 Inches in Diameter, ‘/a Inch High), Open and Assembled Courtesy American Can dompany

(Below) Cans of Commercial Size Used in Experimental Packs, Open and Assembled; Sample Is 2 1 1 / 1 ~ Inches in Diameter and 4 Inches High

Fortunately we do not have to deal with an infinite quantity of material, so there is destruction of bacteria from the standpoint of preservation of canned foods. Figure 6 shows a rate-of-destruction curve plotted on semilogarithmic coordinates. I n Figure 7 the same curve is plotted on paper with linear spacing of lines. The latter is a convenient method of plotting when i t is desired to extend the curve through many cycles, since any major division of the scale may be used to represent a logarithmic cycle. The vertical scale in this method of plotting represents, instead of the actual

isms surviving at the point of destructiom, arbitrarily chosen as that for which slope value a is to be established. The rate-of-destruction curve for the lowest temperature should be chosen as the reference curve. The value of a is allowed to be arbitrary because the expression, zaw- Z log (a/100), gives the time, za, corresponding to a per cent survi.ival, Where this time is of no interest, a may be assigned zz small value such as 1 - 8 , in which case Equation 1 becomes:

100

50

10

5

I 0

50

100

150

xao

250

TIME OF HEATlNG IN MINUTES

Figure 6. W a h o f - esiiruefion Curve on S e m i hagarithrnic CoorddnaPes (Zeta = 68)

number of surviving bacteria, the logarithm of that number. Since the iTertical coordinates are logarithms, the curve is still a straight line, The vertical scale could be expressed in terms of the logarithm of the percentflagof surviving microorganisms without altering the curve essentially. The symbol, Z, for the slope value of the rate of destruction curre was suggested by Raselt. Z is expressed as minutes.

uilcling thermal death time

GUIPV~S

PXIANTOM T H E R ~ ~DEATH ~ A L TIMECURVE. To establish any semilog curve, two items of information are essentiala slope value and the absolute location of one point. In canned food processing, the point usually chosen is that representing 250’ F.(121’ (2.); its value, in terms of minutes, is designated by F . This is regarded as a reference point for all thermal death time curves. If each rate of destruction curve passed through the point representing 100 per cent survival in nil time, the slope value, z, of the thermal death time curve could be obtained from a series of rate-of-destruction curves for a n organism simply by plotting the slope values, 2, of these curves against temperature. Sometimes, however, there is no apparent reduction in number of viable microorganisms for a period of time after heating of the bacteria begins. This period is taken into account in deducing the value of z b y the equation:

One rate-of-destruction curve is chosen as a reference curve, for which T’represents temperature, z& represents the time In minutes during which there is no reduction in the number of microorganisms, and Z is the slope. T,zlJ0, axid Z have corresponding meanings for any rate-of-destruction curve of the series. a is the percentage of original number of organ-

When we determine a value of z by Equation 1 or 2, we have what may be regarded as a thermal death time curve with direction but not position-that is, position with respect to destruction time or end point. Lacking this latter essential. the curve cannot have specificity or definite value in so far as the thermal death time coordinate is concerned. Therefore rye shall call it a “phantom” thermal death time curve. Use of this mathematical procedure associates the idea of incompleteness with the operation, which is desirable since the operation does not establish the real thermal death time curve. CHQICEOF DESTRUCTION PoIxr. To obtain the second essential item of information, which gives position to the thermal death time curve, we must first determine what degree of destruction is necessary; in other words, we must establish a criterion for designating the point at which the ability of the bacteria to decom.pose the food is no longor in-. ciicnt,ecl. We may ask ourselves: “Shoulcl the required degree of dest,ruction be indicated by two remaining viable organisms per unit of ma,terial, one organism per unit, one organism per two units, one organism per ten units, or :ihould it be 0.01 per cent of the number present before heating, as some investigators have suggested?” The a n s ~ ~ to c r this qucstion will he unique for each organism under every different sct’ of conditions. It will depend upon three factors---namely, the nature of the rnicroorgai~isin,the nature of the food under consideration, and the number of organisms originally present in the Lood. To obtain the correct ansTver through laboratory tests, clearly one must give careful consideration t o three additional factors-namely, the temperature of culturing, the culture medium employed, and the number of organisms originally present in the test suspension. The use of a container, whether it be tube or c:m, 8,s tho unit to be counted in establishing results has been criticized by some investigators, because the last spores t o remain viable in a run do riot germinate readily under certain conditions of subculturing or of direct culturing. Such criticism. indicates that the culturing technique is an important factor in viability tests, and that this techniq,ue must be chosen according to what the object of the test is. It appears that the difficulty in producing growth in the last fen organisms to survive a heat treatment may be caused by the fact that the severe heat treatment has deprived these organisms of a portion of their vitality. The condition may be purely a phenomenon of delayed germination. The difference between the effect of this phenomenon upon the results of subculturing in peptone-dextrose broth, for example, and the effect upon the results of plating in tryptonedextrose broth and agar is no doubt one of degree only. It may be unreasonable to expect that all bacteria-still vital, although injured by heat---will produce colonies on agar plates. Therefore, under any method of testing there may always be some vital spores remaining wvhich cannot germinate within the time usually allowed in laboratory determinations. MATINGOF BACTERIA. If one wishes to give free rein t o his imagination, he may conclude that the apparent lack of

INDUSTRIAL AND ENGINEERING CHEMISTRY

January, 1943

vitality in tests by laboratory culturing technique, of the last few remaining organisms which our curves indicate are still vital, is accounted for by the existence of a sex characteristic in bacteria, and that with the organism so sparsely distributed, opportunity for mating is not afforded. Of course, there are many known facts regarding the propagation of bacteria

I,ooo,ooc ORGANISM 1518

(TEMP. OF HEATIN6=220aE)

I00,0oc

:

TEST BY NATIONAL C A N N E ~ S~ 5 5 ' ~ .

gl0,oOc

I IOOC

z

a

c

Ioc

K

IC

P z

liI

2

; a

0,

m

0.0

onc I50 100

0

100 200 300 4W TIME OF HEATING IN MINUTES

500

Figure 7 . Rate-of-Destruction Curve on Linear Coordinates

77

will not be discussed except for the statement that a logical laboratory test to establish a basis for choice of a destruction point is t o culture portions of material containing very small numbers of vital spores, taken during the latter part of a heat resistance run, in food identical to that for which processes are being studied. The results of such culture tests give a good indication of what degree of destruction might reasonably be considered the destruction point, or the disabling point, of the spores in that particular food. Even after this has been done, the thermal death time will still be shown only conditionally by rate-of-destruction curves. We shall have found the end point, but the beginning point will then have to be considered. What number of spores per unit of food material did we have originally? The rate-of-destruction curve shows clearly that the smaller the original number of bacteria, the shorter the time required to reach the end point unless the end point is expressed as a percentage of the original number. For instance, Figures 6 and 7 show that there were about 21,000 spores in each unit of material a t the beginning of the test represented by the rate of destruction curve. The time scale to the left of the zero point in Figure 7 indicates the additional time that would be required to reach any given degree of destruction if the number of spores originally present were greater than 21,000 per unit of material. For example, if there had been 1,000,000 spores per unit of material, almost 125 minutes longer would have been required to destroy them than was necessary with 21,000 spores per unit of material. This effect of numbers upon destruction of bacteria constitutes additional objective knowledge which cannot be explained by known facts pertaining to the nature of bacteria. Why should animate cells, in their reaction to heat, exhibit properties similar to those of inanimate molecules in a chemical reaction? The answer to this question must lie in the fact that the reactions that destroy bacteria are chemical.

Experience in processing which discredit such a theory. Efforts of Sherman and Wing (24) to reveal sex in bacteria were futile; nevertheless the fact that the possibility of mating is not to be discarded too lightly is impressed on those who read Jennings' report (12) of his pryings into the private lives of certain single-cell organisms, the ciliate infusorians. Jennings considers the plausibility of the mating of these 'organisms, and his discussion of under-par individuals indicates a possible explanation of the apparent low virility of spores that have been severely heated. Some may be able to germinate after a long period; others may never be able to germinate even though presumably they are still vital. On this hypothesis a n adequate sterilizing process may be defined as one that will carry bacterial flora into the stage in which any vital bacteria that remain will not be able to germinate. TESTSTO DETERMINE ENDPOINT.Viewing the situation in the light of these circumstances, one must consider carefully the question of what is the best practicable test to reveal the destruction points of spores. Since any destruction point based on the results of either laboratory or practical test must be, at best, an arbitrary point, the choice of a method to determine such a point should depend upon which arbitrary point appears to bear the most consistent relationship to the conditions existing in actual commercial processes. Logically, it is necessary to assume that a similar condition of delayed germination may exist in cans packed commercially to that existing in tubes and plates in the laboratory. Furthermore, if we are to believe what our curves tell us, that some small percentage of the bacteria will always remain vital, perhaps we should assume that the last few vital organisms will never germinate. Several avenues of attack on this problem are open. They

The implication carried by the conclusion, reached on the basis of theoretical analysis, that foods originally infected with microorganisms can never be completely freed from them by sterilizing processes would be disturbing were it not for practical experience. For the last fifteen years, during which processes have been established with the use of thermal death time results that were based on the assumption of complete destruction of microorganisms, experience has validated that assumption from a practical viewpoint. At any rate, it has shown that if vital organisms do remain in the food after processing, they are in such a state that they can do no harm, During this period at least a hundred billion cans of food susceptible to spoilage by Cl. botulinum have been placed on the market. Surely there were spores of C1. botulinum in a portion of that food when it was canned. Reasoning from the indications of rate-of-destruction curves, we should conclude that some of these spores in a hundred billion cans survived the sterilization process. Perhaps they did; nevertheless, they produced no evidence of their survival.

Taken-for-granted safety Circumstances of this nature apply to food of many types. Hypothetical speculation leads us to a realization of the existence of many possible bacterial hazards in our everyday foods. Experience in the common use of these foods without ill effects has inured us to this possibility; apparently the hazards do not actually exist even though we cannot explain fully why they do not. For example, since Cl. botulinum is a soil organism widely distributed in the United States, soil particles on many root and tuber vegetables prepared in the home, such as carrots, turnips, and potatoes, must be in-

78

XNDUSTHIAL AND ENGINEERING CNEM ISX R Y

Vol. 35, No. 1

accustomed, in which there was an element of safety apparently not present in the ne\$-. Xoulton (1.4) attributes tlie contamination of the hams with the organism to carelessness on the part of the packer, retailer, or consumer. Contamination niay have been no greater than in the oltl style product. Perhaps the organisms simply had a better chance to develop than formerly, due to a different chemical environment. I n 1938 a small amount of trouble \\-asexperienced because of the growth of two highly heat-resistant strains of mold in No. 1.0 cans of blueberries. For destruction of the spores, JVilliarns et al. (29) found that a heat treat,rnent of approximately 10 minutes at 200' F. 17-as necessary. It was evident that the spores n-ere not destroyed by the sterilizing process ordinarily used for the product. The question immediately arose as t o why these organisms had not caused trouble before. 'The :~ns117cr seems to lie in the factor of oxygen in the can. Both strains are capable of growing in vacuum of 26 inches. Thk indicates that the amoiint of oxygen required for thcir growth is cxtremely small. It appcars that even this small amount of oxygen was not available in the processed cans when so-called plain cans (no enamel linings) were used. T\lien a change was r u d e 'ro enamel-lined cans, howcver, the absorption of oxygen within the caris as retarded because tlic covering of the catalytic metal was sufficient to permit growth of the mold in the can for a limited period. This explanation i s only teiitntivc, and tlict matter is receiving further study. Figure 8. Thermal Death T i m e Curves of Pathogens, and Phuntom The mmiing of science to industry Irorii the Curves Used in Calculufians hain and blueberry c:sperienccs is t'li:tt the effects of taken-for-granted factors slioi~lrlbe recognized a s early a s possiblc, evcn tjltoiigh these efiecb are not understood. nntl t,liat changes of proc:cfected xith spores of this organistn. Klien these foods are dure or of ;issociated condition-: ~v1iic.hniight hare a ~ignifiwnt prepared in cert,ain v - a y s , anaerobic coiitlitions n-ill exist and influenre on taken-for-granted Factors : L ~ Cto bc :tvoitlctl [ u l l c s ~ presumably provide satisfactory enrironmeiit Cor normal they are :trcoinpariietl by 1ne:rsurcs t o couiitemct ally ~.ctiuc>groq-th of the spores n-hiie the prepared foods are held t,ioii in snfety factors that' t,hechanges m i g h t ~~rotlucc. before use. Still, recolds fail t o report a single case of bot'uR I E N T OF TP;CHKIQUJG, This se:rrcliing loi, I>ettcr lism froni such foods. methods of finding the destruction points of 1i:ictcria is nior Consider even pasteurized nrilk. T2at'A-of..destructiol? a part of a process of refinement of tecliiiique for tire purpo principles would convince us that, even in porfcctly pasteurof placing our attack on processing prcil~lcriionto :i1)ro:iti ized milk, if the amount of milk is great, some vital nonsporeplane, so :is t o bring into consitleriition inore and I ~ O I ' C of : tile forming pathogens n-ill remain. Experience feachcs 11s th:tt, factors :tiit1 t o put o ~ i rfindings onto :iiiiorc (Miin effect, this is not true. In other ~r-ords,w e are taking s t e l i s Lo : i i i ~ ~ : i . These are samples of "taken-for-graneet~" safety in foods, "\\'h$?". as applied b o matters i,h:tt h r c 11:itl t o safety which our most, advanced theoretical knovledge tries he taken for gimited. to convince us does not exist, bi.it which experience, suppleThe loaical. expectation lrorn thc a1)plic:it'ion o f tlic mtt:-. mented by actual scientific tests, shows does exist. JTe mist of-dcstruct,ion principle to t'hermal tle:it,li tjnic tietorxiiiri:itio~i he cautious, however, lest n'e become so accustomed to taking is that, when w e succeed i n rrixliirig it ivork, \ve siltlli Iiave for granted the safet,y of processed foods that. we forget to linowledge of the control of I-iactcria, that rcill pcmmit I X H t,o apply adequate safeguards xhen changes in processing proreduce proresses .ii.ith safety from thosc n o r rega cedure are made. essa,ry. There must be mimy oombinations of condjtiorrs The ordinary boiled aiicl baked hams of commerce have not yet understood, under which organisms will riot, been accepted without question for years, and have never gr o JV . been responsible for a case of gastrointestinal irritation froin bacterial cause except when the ham was grossly mishandled. Yet during 1940 seventeen cases of food poisoning were traced RZillr pasteurization processes to organisms of t,he staphylococcus group occurring in tenMosl; of the problems associated with milk pasteurixation dered hams. This is a small nuniber of cases compared t o t h ~ : can be solved by application of tlie principles just espl:Lineti: number of hams consumed; but it is enough to indicate that notwithstanding the uncertainty that still exists in rcspect tendered ham may be more susceptible t o growth of this ort o some of them. Rate-of-rlestructioli c u r w s for thc ganism than were the types of hams to which we were formerly

January, 1943

INDUSTRIAL AND ENGINEERING CHEMISTRY

non-spore-forming pathogens yield phantom thermal death time curves which, when combined with time-temperature heating, holding, and cooling curves for milk in the pasteurization processes, enable direct comparisons of lethal values of different pasteurization processes to be made with respect to individual organisms, both pathogenic and thermoduric. Fortunately, experience often comes to our rescue when we seem to be stopped on the theoretical path. I n the present case experience in pasteurization provides the information required for establishing real thermal death time curves after phantom curves have been obtained from results of rate-ofdestruction tests. If one desires only to calculate a process that is equivalent in lethal value with respect to a specified microorganism to a given process which is known to be adequate to destroy that organism, the slope is all that needs to be known about the thermal death time curve. I n other words, the phantom thermal death time curve, together with one process that is satisfactory in practice, is sufficent to establish the real thermal death time curve for the microorganism and, from this, to establish other processes. Thus, a given process serves the same purpose, in establishing the position of the thermal death time curve, as an end point of destruction would serve if determined by thermal death time tests in the laboratory. Standard calculation procedure provides a simple way of finding the F value of a curve when the heat penetration curve and the phantom thermal death time curve are known. There are statements in the literature (25, $6) to the effect that pathogens are destroyed to an equal extent by the highshort process of 16 seconds a t 71.7" C. (161" F.) as by the but that long-hold process of 30 minutes a t 61.7" C. (143" F.), this is not true with respect t o thermoduric bacteria. This

79

means that slope values, z, of thermal death time curves for pathogens are lower than those for thermodurics. The simplest explanation of the significance of variations in z is that the lowest z value implies the highest relative sterilizing value for a high temperature as compared to that of a low temperature. For example, if two microorganisms have thermal death time of 30 minutes a t 61.7" C. (143" F.), but the thermal death time curve for one has a slope of 18" F. (10" C.) whereas that of the other has a slope of 9" F. (5" C.), the sterilizing value of 1 minute a t 71.7" C. (161" 17.) will be ten times as great with respect to the latter organism as to the former. If we examine pasteurization specifications of 16 seconds a t 71.7" C. and 30 minutes a t 61.7" C. in the light of the thermal death time curve principle, we find that any organism for which these two processes hold equal destruction value has a thermal death time slope value of only 8.7" F. (4.8" C.). This is indicated by curve X on Figure 8, which passes through points P and Q representing, respectively, the two specified processes. This statement assumes that no time is required to heat the milk to the set temperature of either 71.7" or 61.7" C. or to cool it a t the end of the process; that is, the milk in every case is brought to the pasteurizing temperature instantaneously and is cooled instantaneously. Such an assumption is contrary to fact. I n all cases time is required to heat the milk to the designated holding temperature for pasteurization. To take this into account in a study of pasteurization by the method used for canned food sterilization, one must indicate the temperature of the milk during its rise to the holding point by a curve which will occupy a place in this problem analogous to that of the heat penetration curve in problems on canned food processing. Similarly, one must represent the rate of cooling by a time-temperature curve.

Courtesy, Cherry-Burrell Corporation

Milk Pasteurizer of Short-Hold Type, Plate Construction

The ascribing of comparatively low z values to thermal death time curves for microorganisms of low heat resistance is consistent with practice in canning process studies. I n proeess calculations t o cover destruction of nonspore formers in tomatoes, for example, a 2 value of approximately 14" F. (7.8" C.) has been considered correct, as compared with 18" F. (lo* C.) which is most commonly used for spore formers in nonacid products. Williams et al. (29)found thermal death time curves for the faculative anaerobic mold in blueberry juice, previously described, to have a slope value of about 10.5" 2'. (5.8" C.) Our studies with nonacid foods have never carried us into the use of a values in the extremely low region that seems t o be callled for in high-short pasteurization Work.

Table I.

Lethality Constants of Heating and Cooling ]Poreions a% &%ilk hstelarizatismn $recess

z Value

C.

F.

Constmt A m o r A,

Constant B

Constsnt C

Constent D

53.6 56.6 59.3 61.9 64.3

2.17 2.36 2.56 2.76 2.95

3.15 3.34 3.54 3.73 3.93 4.13 4.32 4.53 4.71 4.91 5.10

6.11 6.67 7.22 7.78 8.33

11 12 13 14 15

466 510 554 599 644

144 148

8.89 9.44 10.0 10.5 11,l

16 17 18 19 20

688 732 777 821 866

161 163 166 171

66.6 67.8 70.6 72.4 74.2

11.7 12.2 12.8 13.3 13.9 14.4

21 22 23 24 25 26

910 955 1000 1044 1089 1133

173 I75 177 179 181 183

75.8 77.4 78.9 80.4 81.3 83.3

152

155 158

(Figure 9) represents heating during 7 minutes, curve Acl, cooling during 7 minutes. Since heating is by indirect mean5 in milk pasteurization, the rates of heating and cooling ordinarily are not expressible as simple linear functions. They seem more nearly to resenible semilogarithmic functions. Factors B, C, and D apply to semilogarithmic Pates in various modifications. B factors of Table I apply t o a heating rate represented by a straight line on semilog paper (Figure 91, Since such a curve would go to infinity unless it were defined, a termination a t some point below holding temperature must be specified in order t o establish a slope value for a rise of temperature within a stated time, B factors were determined on the basis of terminating the rise of temperature 0.056" C. (0.1" I?.) below the holding temperature, H T . G factors are analogous t o B except that the curve is made to terminate 0.56" C. (1.0" F.) below holding temperature, HT. They apply to the heating period only (Figure 9). D factors apply t o 8 cooling rate represented by a straight line on semilog paper (Figure 9) the terminating point of which is supplied optionally for each calculation. The terminating point is expressed as the number of degrees above the temperature of the cooling medium, G W , and is represented by ge for Fahrenheit and 0.56 gofor centigrade. A lag factor, commonly Pepresented by. j, is essential in semilogarithmic heating and cooling curves. Factors B , C, and D are based u p ~ nj = 1.0, which seems reasonable for convection heating of a liquid product in mechanical agitation. Knowledge of this value is not essential t Q carrying out these calculations. The calculation of the percentage values €or lethal effect during the periods of niaing and declining temperature are made by Equations 3 to 7:

A for cooling period: p,a =

By the use of arbitrary constants, the lethal value of the temperature rising and declining periods of the pasteurization process can be readily calculated, based on assumed rates of rise and decline of temperature. Table I contains lethality constants A , B , C, and D, respectively, for four different rates of rise and decline of temperature. Simple calculations will convert the constants into percentage values referred t o the amount of lethal heat set as the requirement of the process. For example, if 30 minutes a t 61.7" C. is regarded as the measure of sterilization required for a given microorganism, conversion of the constants in Table I will give the amount of lethal heat that is effective during the periods of temperature rise or decline, expressed as a percentage of the amount required for the process. Assuming that the phantom thermal death time curve for the microorganism is known, a thermal death time curve having the slope value of the phantom curve and passing bhrough the point representing 30 minutes a t 61.7" C. will show for all temperatures the times required to give sterilizing effect equivalent to heating for 30 minutes a t 61.7" C. when both the rise to and decline from holding temperature are instantaneous. The percentage values yielded by the constants of Table P apply to the period of rise to or decline from the holding temperature, regardless of what that temperature may be. A factors in Table I apply to either a uniform rate of rise from initial temperature I T to holding temperature H T , or of decline from holding temperature BST to final temperature PT-that is, a change through a constant number of degrees during each successive unit of time. For example, curve A R ~

R for heating period: p s

=

C for heating period: p c =

A L A

U(HT

- PT) BtRR

-+ 11

U [ l o g(MT - IT) CtRC

U log (HT

('I

-IT)

where P E A , p , ~p~ . p c , p~ = % of required lethal heat AB,A,, B,6, D arbitrary constants from Table I H T = holding temp. of pasteurization process I T = initial temp. of milk FT = final temp. of milk after cooling CW = temp. of cooling medium t R A , t ~ stsc , = time consumed in rise of temp, of milk, min. ted, t,a = time consumed in decline of temp. of anilk, nnin?. U = time necessary t o destroy microorganism at holding temp., min. difference between PT and CW, degrees go i=

The time during which the milk must be heid a t holding temperature NT in order t o give adequate pasteurization when time is consumed in the rise and decline of temperature of the milk for different combinations of rates of heating and cooling is given in Equations 8 to 13:

INDUSTRIAL AND ENGINEERING CHEMISTRY

January, 1943

0.01 A f i t R . 4 HT - I T

-

0.01

81

temperatures in a single series of determinations was given. The curves are described in Table 11. (10) 0.01 &,A Rate of heating B These data were taken from HT - F T log(HT - I T ) 1 Rate of cooling A Hammer (IO). The rate of heating or of cooling was not stated. 0.01 B t R B 0.01 D f , D for any of the destruction points log (HT - Cw) - log log (HT - I T ) 1 Rate of cooling D given, nor were the intervals of time between readings for four of Rate of heating C - 0.01 A L t C A (12) the five tests. The concentra- 0.01 C t R C HT - F T log (HT - I T ) Rate of cooling A t R c A tion of microorganisms was stated for Br. suis only. When time 0.01 C t R C 0.01 D ~ , D (13) was stated somewhat indefilog (HT - I T ) log (HT - CW) - log ~c nitelv-for examde. "less than Rate of cooling D 7 minutes"-the exact time of destruction had to be estimated. where ~ X A At H A D , ~ H B A ,~ H B D ,~ H C A ,t H C D = time milk is held at An important feature of these data lies in the fact that some hofding temperature, HT, under different combinations of bacteria probably were destroyed while the temperature was rates of heating and cooling, min. rising, but since the rate of rise is not given, the absolute lethal value of the nominal process is not revealed. The second and third terms in the right-hand member of The number of minutes indicated by the thermal death each equation express lethal values of heating and cooling time curve of each organism as the resistance a t 60' C. periods, respectively, as minutes of heating at the holding (140' F.) is given in Table I1 under the heading PUlao,a symbol temperature, H T . For example, if in Equation 13 0.1 CtRC/ invented by Benjamin for "pasteurization unit". Designab log (HT--IT) equals 2.5, this fact would signify that the ing a point on a curve, this value serves the same purpose for coming-up period is equivalent in sterilizing value to 2.5 reference and placement of the curve in the pasteurization minutes a t holding temperature HT. temperature range as F does for curves in canned foods procDATAIN DESTRUCTION OF BACTERIA.Most of the data in ess range. the literature are of such form as to be of no value in establishOther data in the literature on the nonspore-forming bacing thermal death time curves. They usually indicate only teria include information on resistance of five types of such isolated points, and controlling conditions, such as concenbacteria in phosphate solution and broth. Beamer and Tantration of organisms and rate of rise of temperature, are not ner (6) and Baker and McClung (1) published rate of destrucspecified. The following quotation from Hammer (10, page tion curves for such organisms. The former included Eber136) is typical of the presentation of these data: thella typhosa, Salmonella paratyphi, Salmonella aertrycke, Salmonella enteritidis, and Staphylococcus aureus in broth of Rosenau found that E. typhosa was killed ip milk heated to p H 7.05; the latter studied Escherichia coli in phosphate soluGO" C. (140' F.) and maintained at this temperature for 2 mintion of p H 7.00. utes; the great majority of the organisms was killed by the time the milk reached 59" C. (138.2" F.), and only a few survived at Slope values of phantom thermal death time curves from 60' C. C. diphtheriae often failed t o grow after the milk reached these data are given in Table 111. Those for Eberthella 55" C. (131' F.), although occasionally it survived until the milk typhosa and Salmonella aertrycke and the first of those for reached 60" C. The resistance of V . comma was similar t o that Escherichia coli are rather clearly defined by the values taken of C. diphtheriae; it was usually destroyed when the milk reached 55' C. but once it lived until 60" C. was reached. from the data while those for the other organisms are indefinite because the three values taken from the data for Figure 8 shows five thermal death time curves plotted from each type of bacterium do not indicate a straight line. the only data found in the literature on pathogenic organisms CALCULATION OF LETHALITY OF PASTEURIZATION PROCin milk for which a destruction point a t each of three or more ESSES. Neither Table I1 nor I11 contains data on bacteria of Rate of cooling D

Dt,D

log(HT - CW) - log g,

+

+

1

I

,

Table 11. Thermal Death Time Curve Data for Pathogens in Milk (10) Curve No.

1 2 3 4 5

Microorganism E.typhosa M. tuberculosis M. tuberculosis

Medium in Which Heated Cream Cream Milk Milk Milk

No. per M1. Not given Not given Not given Not given 500,000,000

{ gi: ~ $1 ~ ~ , Br. ~ u i s

No. of Points Given 3 4

C. 8.5

O

7.0 6.9 8.0 5.1

S 3 3

O F. 15.3 12.6 12.4 14.4 9.2

PUMO 4.5 9.2 14.0

Reference

8.6

(18) (10)

23.0

{if]

(1 7 )

Table 111. Data on Nonspore-Forming Bacteria Bacterium Eberthella typhosa ( 6 ) Salmonella parathyphl ( 6 ) Salmonella aertrycke (6) Saimonella enteritidis (6) Staphylococcus aureus ( 6 ) Escherichia ooli (f) Escherichia coli ( 1 )

Medium in Which Heated Broth Broth Broth Broth Broth Phosphate Phosphate

amocmoni

pH 7.05 7.05 7.05 7.05 7.05 7.00 7.00

51.7'C. (125O F.)

.. ..

62

54.4' C. 55'C. 57.2OC. 60°C. 62.8' C. (130" F.) (131' F.) (135' F.) (140' F.) (145' F.) 8.6 ... 40 ... ... 32 ... 10 13 .. ... 40 ... ... 46 ... 13 .. 33 ... 300 ... 350 158 52 23 13 6

.. ..

... ...

..

.. ..

65' C. (149' F.) 1.6 1.7 4.3 6.4 10 *. ..

zVa'ue 7.3 8.6 10.5 10.9 6.1 6.6 5.1

13.1 15.5 18.9 19.6 11.0 11.9 9.2

82

INDUSTRIAL AND ENGINEERING CHEMISTRY

Vol.

35, No. I.

lethality values of the processes, 1.10 calculations were made; the results are shown in Tables XV arid V. Rmting and cooling t h c s of 0, 7 , and 15 minutes were ai;sxrmcd with dl combinations of heating rates 112, B,and C' xvitb cooling m t o s & and 1 9 (Tiible I). 'Pht? heating time trr required to give, iii corribirrntion with. heating and cooling periods, lethal eft"cct,sequivaJent in t,hi)s(> of px~cesseswith instant,aneoos heating arid conling tts intfjcated by the phantom thermal death hime cur'vcs Tvcre i d ciilnted by Equatjoiis 8 to 13. T o give Lin:tlogous coinp iri other terms, the per cent p of adequate letha suppiitid b y each process Tvheii 111 = ii was calc thus indicating clearly how much lethal heat is accourrtoci for by the corning-up and cooling periods, When tir =: (I, the holding period, hr, aceourits f o y lethal heat c:qual to 100 per cent of that required to pasteurizo. l'ha lat;Lrr computations were made by evaluating the tightliantl 1n01n-~ bels of Equations 8 to 1 3 after changing the signs between terms from minus to plus, dividing the result by the value of I / , and multiplying by 100.

5

IO

Figure 9. Hypothetical H e a f i n g and CooIing Curves for

Milk in Pasteurization Process

thermoduric type, therefore it is impossible to state definitely that thermal death time curves for thermodurics in milk have greater slope values than those for pathogens in milk. Most spore formers in nonacid media have thermal death time curves with slope values in the general vicinity of 10" C. (18' F.) and some have higher values. Assuming that a typical thermoduric microorganism in milk has the same z value as that shown in Table 111 for Salmonella aertrycke-namely, show how processes a t 10.5" C. (18.9" F.)-calculations 71.7' C. (161" F.) that are equivalent with respect to this organism to a given process a t 61.7" C. (143' F.) compare with processes a t the former temperature that are equivalent with respect to BY. si& to the same given process at 61.7" C. Phantom thermal death time curves for these two organisms, passing through point P representing 30 minutes a t 61.7" C., are indicated by lines X and Y on Figure 8. The intersections of these lines with the temperature coordinate representing 71.7" C. at points T and V give the holding times at 71.7" C. which are equivalent, respectively, to 30 minutes at 61.7" C. with instantaneous heating and cooling in all cases. These times are 19.2 seconds for Br. suis and 3 minutes 20.4 seconds for S.aertrycke. To illustrate the effects of various rates of heating and cooling, in conjunction with variation in z value, upon the

Discussion o f results The outstanding fact revealed by thew calculations is that when G has a value of 3'10 minutes or loss, the heat applied during corninkup and cooling periods may be a major portion of that required for. pasteurization. Tlie estreme condition shown is in the pro 71.7" C. (161" I?.). A t this temperature 19.2 seconds arc sufficient to destroy .Br. Sui?, and if 15 minutes mcre used to bring up the temperature at beating rate 23, the heating period alone would supply almost twenty-one tirnes as much lethal heat as is needed to destroy the organism. I n ft process in which p = 200, the leLhal heat of the comingup and cooling periods combined is just sufficient to destroy the microorganism; therefore the required holding time i s nil. When p > 200, 1~ < 0, as indicated in Table V for numerous processes a t 71.7" C. The latter condition signifies that pasteurization is accomplished without heating the milk all the way to HT. The maximum amount o€ lethal heat contributed by a corning-up time of 15 minutes when T/ = 30 minutes is 28 per cent (process 16, Table IV) and the maximum for comingup and cooling periods combined (each 15 minutes long) is 34 per cent (process 17 for S. aertrycke, Table V). The cooling period has little value compared t o that of the coming-up period except when both rise and decline of temperature occur a t rate A , in which case the values of &hetwo periods are identical. The: maximum length of the coming-up and cooling periods-namely, 15 minutes-chosen for these calculations is great compared to the length commonly used in practice. s V) Kevertheless, inspection of the p values for Br. ~ u , i(Table makes it clear that, even though the period were only 1 to 5 minutes in length, its value would be appreciable compared to that of the holding period. I n the interpretation of these results, it is essential to keep in mind that, although they are based upon phantom thermal death time curves, they give accurate comparisons b~tweain different processes a t the same or at different processing temperatures. Table V indicates sterilizing values for processes a t 71.7" C. (161* F.) which are strictly comparable to the values given in Table IV for processes a t 61.7" C. (1463" I?.) for Rr. suis and S. aertrycke as they are represented by available heat resistance data. If the processes a t 71.7' C. a 1 ~ longer than necessary, it is because the process of 30 minutes a t 61.7' C. is correspondingly long. The fact that the coming-up periods of short-time pasteurization processes in practice customarily have not been taken into account in the experimental evaluation r ~ f

January, 1943

INDUSTRIAL AND ENGINEERING CHEMISTRY

I

Table IV.

Lethality Values of Processes a t 143' F. (61.7' C.)

which is that the rate of reaction doubles for each increase of temperature of 10" C. or 18" F. The time-temDerature curve of inactivation will have slope value that is unique for the conditions under which the inactivation is accomplished. On the simple monomolecular reaction principle the slope value would be 33.2" C. or 59.8' F. Whether OF not this rate obtains in milk may not yet have been determined. We shall refer t o this curve as the "inactivation time curve". A phantom inactivation time curve will result from a series of rate-of-inactivation curves a t different temperatures. This curve will establish between different pasteurization processes relationships that correspond completely to those discussed for the curves shown on Figure 8. It will tell what process at 71.7" C. will accomplish the same degree of inactivation as the process of 30 minutes a t 61.7" C. This process a t 71.7" C. will not be the same as those a t the same temperature which are equivalent to 30 minutes at 61.7' C. with respect to microorganisms of which the phantom thermal death time curves are different from the inactivation time curve for phosphatase. If the slope value of this curve is 33.2" C. (59.8" F.), 15 minutes a t 71.7" C. will be equivalent to 30 minutes a t 61.7" C. Thus, if processes are based upon an organism of which the phantom thermal death time curve

-

(IT = 40° F. or 4.4" C ' Q~ = l o F. or 0.56O C.; U = 30; t~ for each process when just adequate for pasteurization; p in per cent of adequate lethal heat in process when t H U) Br. suis S. aertrycke R a t e of R a t e of z = 9.2' F. z = 18.9' F. Temp. Temp. Process FT, cw, F. F. Decline Rise No. tH P to tH P tR 30.00 100.0 40 30.00 0 100.0 1 0 .. 29.44 101.9 28.81 104.0 40 15 A 2 40 A . . 29.74 101.0 29.44 101.9 7 3 41 D 40 29.86 15 100.5 29.73 101.0 4 41 40 D 29.94 7 100.2 5 29.87 100.5 29.44 A 0 40 101.9 6 15 28.81 104.0 40 A .. 28.88 15 103.8 27.62 108.0 7 40 A 29.18 7 102.8 28.25 105.9 8 D 29.30 41 15 40 102.3 9 28.68 28.55 104.4 104.8 41 D 29.38 40 7 102.1 10

..

..

..

..

..

a

..

11 12 13 14 15

7

16 17 18 19 20

15

21 22 23 24 25

7

26 27 28 29 30

15

31 32 33 34 35

7

-4

B

B

c ,

C

..

0 15 7 15 7

A A D

0 15 7 15 7

A A D D

0 15 7 15 7

A A D D

0 15 7 15 7

A A

0 15 7 15 7

n

.. ..

..

n n .

I

A

A D D

40 40 40 41 41 40 40 40 41 41 40 40 40 41 41 40 40 40 41 41 40 40 40 41 41

..

.. ..

29.74 29.18 29.48 29.60 29.68

101.0 102.8 101.8 101.3 101.1

.. .. ..

23.27 22.71 23.01 23.13 23.21

40 40

40 40

.. .. I

.

40 40

.. .40...

40

..

.. .. 40

40

the processes may be responsible for the fact that an undue amount of credit seems to have been attributed to the temperature of 71.7" C., in considering that 16 seconds a t that temperature are equivalept t o 30 minutes a t 61.7" C. A part of the lethal effect attributed to the holding time a t 71.7' C. perhaps rightly belongs to the coming-up period. If so, a slope value greater than that of curve X, Figure 8, is indicated, It is true that the slope value indicated by data for Br. suis is very close to that of curve S. The available curve for Br. suis, however, is one of the less well defined curves. The three points given do not indicate it clearly. An additional fact bearing on the comparisons is that the data on 8. aertrycke were made during a heat resistance test with the use of broth instead of milk. Phosphatase test Just as bacteria of a given strain are destroyed by heat in accordance with a timetemperature relationship shown by a thermal death time curve, the enzyme phosphatase undoubtedly is inactivated by heat in harmony with a somewhat similar pattern. Since phosphatase inactivation probably is more nearly a pure chemical reaction than is the destruction of bacteria, the reaction under certain conditions might proceed on the simple monomolecular reaction principle,

122.4 124.2 123.3 123.0 122.6

29.44 28.25 29.18 28.89 29.37 21.63 20.44 21.07 21.36 21.50

101.9 105.9 102.9 103.7 102.3 128.0 134.0 129.8 129.0 128.4

26.87 26.31 26.61 26.73 26.86

110.4 112.3 111.2 111.0 110.4

26.06 24.87 25.50 25.93 25.79

113.2 117.1 115.0 113.7 114.0

26.48 25.92 26.22 26.34 26.42

111.7 113.7 112.7 112.3 112.0

28.40 27.84 28.14 28.26 28.39

105.3 107.2 106.2 105.8 105.3

24.61 23.42 24.05 24.34 24.48 27.49

118.0 122.0 119.9 119.0 118.3 108.3

26.30 26.93 27.22 27.36

1112.3 10.3 109.3 108.8

'

Table V.

83

Lethality Values of Processes at 161' F. (71.1' C.)

.

(IT = 40' F. or 4.4" C . :

Q~ = 1' F. or 0 56' C &for each process when just adequate for pasteurization agamst the organism; p (n per oknt of adequate lethal heat in process when 1H = U) Br. suis, 8. aertrycke, z = 9.2' F. E = 18.9' F., Rate of R a t e of U = 0.32 u = 3.34 Process Temp. Temp. cw, No. Rise Decline 2R tc F. tH P tH P 1 0 .. 0 .. 40 0.32 100.0 3.34 100.0 2 15 A *. 249.0 2.32 40